Cancer in Florida Hispanics

Transcription

1 FLORIDA CANCER DATA SYSTEM SYLVESTER COMPREHENSIVE CANCER CENTER Cancer in Florida s 99 Monique Hernandez Lora Fleming Jill MacKinnon David Lee U NIVERSITY OF MIAMI M ILLER S CHOOL OF M EDICINE

2 CANCER IN FLORIDA HISPANICS 99 Monique N Hernandez, PhD Lora E Fleming, MD PhD MPH MSc Jill A MacKinnon, PhD CTR David J Lee, PhD Florida Cancer Data System University of Miami Miller School of Medicine Miami, Florida i

3 ACKNOWLEDGEMENTS The authors would like to thank the staff of the FCDS and the Cancer Registrars of the State of Florida for their contributions to this Monograph, as well as Dr. Paulo Pinheiro MD PhD for his unique contributions to the study of cancer. The Florida cancer incidence data used in this report were collected by the Florida Cancer Data System under contract with the Florida Department of Health (DOH) and the National Program of Cancer Registries (NPCR) of the Centers for Disease Control and Prevention (CDC). The views expressed herein are solely those of the authors and do not necessarily reflect those of the CDC or the Florida DOH. Telephone () - Florida Cancer Data System website Suggested citation: Hernandez MN, Fleming LE, MacKinnon JA, Lee DJ. Cancer in Florida s 99-. Miami: Florida Cancer Data System,. This work was supported by the Florida Department of Health (Contract COANF) and the Centers for Disease Control and Prevention through the National Program of Cancer Registries (U/DP-). ii

4 Table of Contents A. INTRODUCTION... B. BACKGROUND... B.. The Paradox... B.. Cancer in s... B... Females... B... Males... B... Children... B... Other Cancers... B.. Cancer Screening and Risk Behaviors... B.. Subpopulations... C. METHODS... C.. Assignment of Ethnicity and Subpopulation... C.. Cancers... C.. Cancer Rates... C.. Joinpoint Analyses... C.. Proportional Incidence Ratios... D. RESULTS... D.. Florida Population and Cancer Cases... D... Florida s by Percentage of County Population... D.. Florida s by Percentage of Total State Population... D... Florida Distribution of Cancer by County... D... Percentage of New Cancer Cases by Ethnicity and Site,... D.. Trends Top Cancers Joinpoint Results... D... All Cancers... D... Lung and Bronchus Cancer... D... Prostate Cancer (males only)... D... Breast Cancer (females only)... D... Colorectal Cancer... D... Bladder Cancer (including in situ)... 7 D..7. Head and Neck Cancer... 9 D... Non Hodgkin s Lymphoma (NHL)... D..9. Melanoma... D... Ovarian Cancer (females only)... iii

5 D... Cervical Cancer (females only)... D... Stomach Cancer... D... Liver Cancer... D... Gall Bladder Cancer... D.. Percentage Distributions for All Cancer... 9 D... s Versus non Whites by Gender... 9 D... Proportional Distribution of Cancers by Subpopulation and Gender... D... Subpopulations by Gender... E. DISCUSSION... 7 E.. Cancer Trends in Florida s... 7 E.. Stage Specific Cancer Trends in Florida s... 7 E.. Comparison of Select Florida Incidence Rates with Pooled Data from State Registries... E.. Cancer Distribution within Subgroups... 9 E.. The Paradox and Cancer in Florida... 9 E.. Study Limitations... E.7. Future Research and Prevention Activities... APPENDIX. Major Cancers: International Classification of Disease (ICD) and FCDS Coding... REFERENCES... iv

6 A. INTRODUCTION In the United States, s are the largest, youngest and fastest-growing minority accounting for % of the US population with. million people in 7. In, there were. million (%) s among Florida s rapidly growing population. Although traditionally the majority of population growth in both the US and Florida has been due to immigration from Spanish-speaking countries, recently births have become the largest source of population growth. Traditionally, s as a group have experienced lower cancer incidence and mortality rates. However, as described below, recent studies have demonstrated significant variation among subpopulations within all US s with regards to their cancer experience. This variation has significant implications for the future cancer burden as well as cancer prevention and control in the US, given the projected rapid and continuing growth of both the immigrant and US-born populations. This Monograph focuses on the cancer experience of Florida s (including subpopulation analyses) from 99 through, including identifying potential health disparities. B. BACKGROUND As noted above, s are the largest, youngest and fastest-growing minority in the United States. Following California and Texas, Florida ranks third among states with the highest number of s, totaling over.7 million in 7. In Florida, s make up roughly % of the total population, composed of 9% Native-Born s and % Foreign-Born s. Cancer in s is an important issue because overall the population in the US has demonstrated lower cancer incidence and mortality rates than non- populations. -7 However, in the US, s are a heterogeneous group, immigrating originally from Mexico, Puerto Rico, Cuba, and other countries in Central and South America. Furthermore, s may also differ significantly by their degree of acculturation and socioeconomic status. For example, acculturation to US lifestyles and behaviors has been associated with an increase in obesity, substance abuse, and less nutritious diets among the US population (as well as positive behaviors such as an increase in reported exercise), particularly with increasing time spent in the US. occurrence and risk factors can vary among s because of acculturation, geographic, behavioral, and/or genetic differences. 7, 9 Therefore, cancer B.. The Paradox Since the 9s, despite lower socio-economic status and lower rates of health insurance coverage, US s as a group have been noted to have better health outcomes on indicators such as infant mortality, life-expectancy, mortality from cardiovascular diseases, mortality from major types of cancer, and measures of functional health. Labeled as the Paradox, various factors have been explored to explain the relative advantages of US s including cultural practices, family supports, selective migration ( healthy immigrant effects ), diet, and genetic heritage, as well as Census undercounting, misclassification of,,, deaths, and emigration. Of note, this Paradox has been found particularly among Mexican Americans (and not necessarily among other subpopulations), and is concentrated among those at lower levels of socioeconomic status, with little or no advantage found at higher socioeconomic levels. B.. Cancer in s In general, US s as a group experience lower cancer incidence and mortality rates compared to US non- Whites and non- Blacks. The leading cases of cancer incidence and cancer death are similar among US s as among non- Whites: breast, lung and colorectal cancers among 7, females, and lung, colorectal and prostate cancers among males. However, certain cancers remain increased in incidence and mortality rates across multiple studies for US s as a group: cancers 7, 9,, - of the stomach, liver, uterine cervix, penis, gall bladder, and acute lymphocytic leukemia. In addition, s are more likely to be diagnosed at a later stage for certain common cancers, and less likely to report 7, utilizing cancer screening.

7 B... Females In multiple studies, US females have been found to have increased rates of cervical, gall bladder, 7,,, stomach, and liver cancer, as well as multiple myeloma. As with other race-ethnic groups, lung carcinoma is increasing among females. B... Males In multiple studies, US males have been found to have elevated rates of liver, stomach, gall bladder, possibly nasal cavity, penis, and possibly thyroid cancers, acute lymphocytic leukemia, and Kaposi s 7, 9, sarcoma. B... Children Although few in number, studies of US children have shown increased rates of leukemia, Hodgkin s, lymphoma, and germ cell tumors. However, differences have been found between different states with varying proportions of subpopulations. For example, the incidence of lymphoma, central nervous system tumors, sympathetic nervous system tumors, and malignant bone tumors was highest among youth in Florida (primarily Cuban and Central American s); the incidence of hepatic tumors was highest among youth in California (primarily Mexican s). B... Other Cancers Although traditionally s have not been thought to be at increased risk for skin cancer, recent evidence suggests that melanoma incidence and particularly mortality may be increasing among s. 7 In part this may be due to lack of knowledge about this risk among both s and their healthcare providers. B.. Cancer Screening and Risk Behaviors, 7 As a group, s tend to have fewer known cancer risk behaviors. These include smoking prevalence rates of % among versus % among non- White adults, while % and 7% of females and males, respectively, report not consuming alcohol (vs. % and % of non- White females and males, respectively). Nevertheless, there is considerable variation among s for some of these cancer risk factors; for example,.9% and.% of Cuban males and females, and.% and.% of Puerto Rican males and females, respectively reportedly smoke. 7 And as with non- Whites, obesity is a rapidly growing issue among s, again with considerable subpopulation variation including % of Mexican American males and females measured as overweight or obese. In general, s are less likely to participate in preventive cancer screening. In part, this may be due to a significant lack of health insurance among s compared to non- Whites and Blacks; beyond access to care, other issues have been identified as barriers to cancer screening including: less, education, lower socio-economic status, and cultural barriers. Among Floridians years or older based on the Behavioral Risk Factor Surveillance Survey (BRFSS), with regards to colorectal cancer screening, only 7.% s compared to 9.% non- Whites reported ever having home fecal blood testing, while only 9.% vs. 7.7% reported ever having had sigmoidoscopy or colonoscopy. Among Floridian females, only.7% of females compared to 7.% of non- White females reported ever having had a mammogram, 7.% vs. 9.% a clinical breast exam, and.% vs. 9.9% a pap smear. B.. Subpopulations When data are available, as noted above, cancer incidence and mortality rates may vary considerably by subpopulation. In, out of the estimated million s in the US, % were of Mexican background, 9% Puerto Rican,.% Cuban,.% Salvadoran and.% Dominican; the rest are of some other Central American, South American or other /Latino origin. In, there were.7 million s among Florida s rapidly growing population of million, with an estimated % Mexican, 9% Puerto Rican, % Cuban, and % other subpopulations. The majority of these Florida s are first generation, having immigrated to the US at different times in their life. However, this diversity among s, generally defined by birthplace or geographic origin, has been neglected when assessing cancer risk.

8 C. METHODS The data in this Monograph were derived from all cancer cases residing in Florida, diagnosed between 99-, and reported to the Florida Cancer Data System (FCDS). The FCDS is a statewide, population-based cancer incidence registry created by the State of Florida Department of Health in 97, and operated by the Sylvester Comprehensive Cancer Center at the University of Miami Leonard M. Miller School of Medicine with support from the Florida Department of Health and from the Centers for Disease Control and Prevention (CDC) National Program for Cancer Registries (NPCR). C.. Assignment of Ethnicity and Subpopulation For the assignment of ethnicity and subpopulation, the recently developed HOIA ( Origin Identification Algorithm) was used, using data from the FCDS. HOIA is largely based on the existing North American Association of Central Cancer Registries (NAACCR) Identification Algorithm (NHIA). 9 HOIA takes into account all information routinely available to cancer registries, and in addition, all non- cases are matched to a surname list. HOIA is available online at and has been described in detail in previous publications. - A comparison between results from HOIA and NHIA has been, performed. In short, HOIA corrects for data miscodes common in the FCDS database in the NAACCR data item 9 Origin, e.g., misclassification of unknown s as Mexican, or the inclusion of Brazilians and Portuguese as s. In addition, HOIA uses a stepwise approach to incorporate the information present in death certificates (birthplace and recorded subgroup) with the same information from cancer registry records. Not only does HOIA provide increased ascertainment of ethnicity, it also allows for estimates of cancer rates in the following subpopulations: Mexican s, Puerto Ricans, Cubans, and New Latinos (all other s). 9 Of note, HOIA was not able to reclassify all s into these subpopulations; in this case, these s were categorized as a subgroup denoted as NOS or not otherwise classified. Of note, the rates presented below were focused on s and the comparison group of non- Whites, i.e., a mixed ethnicity and racial classification. s include both Blacks and Whites in part because this follows the patterns of race-ethnic self-identification (i.e. Black s often identify as s rather than Black ), and because the numbers of identified Black s in the FCDS database are quite small. These analyses do not include non- Blacks who are a mixture of African Americans and Blacks from other countries (particularly the Caribbean). As with all data in the FCDS, these race-ethnic data are extracted from the medical and pathology records by trained Certified Cancer Registrars using nationally recognized standards. C.. Cancers For the incidence and mortality analyses, all records of invasive cancers diagnosed among Florida residents of all races and ethnicities during the year period from 99- were used in the analysis. Primary cancer site and histology data were coded according to the International Classification of Diseases for Oncology edition in use at the time of diagnosis, converted to the third edition, and categorized according to SEER site groups. The top cancers among all Florida residents in the FCDS database for were selected, as well as additional cancers demonstrated to be elevated in populations, as described above in the Background section (Appendix ). C.. Cancer Rates Age and gender-specific population data for the state of Florida for each racial/ethnic group for the study years were obtained from the US Census Bureau for the underlying denominator of all persons at risk. As noted above, for the incidence analysis, all records of invasive cancers diagnosed among Florida residents of all races and ethnicities during the year period were used in the analysis. Cancer incidence rates for years (99 ) per, persons were age-adjusted by age groups (, 9,...,, and above) to the U.S. standard population. Age-adjustment is a process to correct for the differences in cancer cases and death counts caused by differing age composition among different populations and counties. The direct method of age-adjustment was used to calculate age-adjusted incidence and mortality rates in this report. Standard errors and 9% confidence intervals (9% CI) were generated using equations published by SEER*Stat. These values were produced to enable long-term cancer incidence

9 trends (99-) through Joinpoint analysis for all s and non- Whites. To protect confidentiality, data were suppressed when cell counts were less than cancer cases (following FCDS rules). C.. Joinpoint Analyses The analyses of cancer incidence trends between the years 99- were conducted using the Joinpoint regression model, where statistically significant rate changes (increase or decrease) determine the best fitting points, or joinpoints. The analysis begins with a minimum number of joinpoints (e.g. zero or a straight line), and tests whether one or more points are significant and whether they should be added to the model by means of the Monte Carlo Permutation method. The final model represents a statistically significant change in a trend at each joinpoint. The Annual Percent Change (APC), or the average rate of change in a cancer rate, was generated for each joinpoint segment and was tested at the ρ <. to determine if the rate of change was significantly different from zero. The Joinpoint analyses were performed using the Joinpoint software, version., from the Surveillance Research Program of the US National Cancer Institute (available at Of note, two different statistical models are used in the Joinpoint analyses, linear and log linear. In order to generate the APCs, the log linear model is used; however, if there are any zero values in the dependent variable due to small sample sizes in particular subpopulations, the log linear model drops that particular value from the analysis, and neither an APC nor a graphic line is generated by the software. In these cases (e.g. subpopulations of gall bladder cancer), we elected to use the linear model to generate a graphic line which is noted in the affected figures ( C.. Proportional Incidence Ratios An age-standardized proportional incidence ratio (PIR) analysis was performed in lieu of incidence rate ratios due to the lack of detailed census population data for sub-groups. This analysis compares the distribution of cancers by anatomical site between sub-groups to the distribution in non-s. This analysis was derived by calculating the expected number of cancer-specific cases in each sub- group, and the PIR represents a ratio of the cases observed to those expected. For each sub- group, this expected number was obtained by multiplying the total number of cases for each age by the corresponding age-cause-specific proportions in the non- white population. An elevated PIR does not necessarily indicate that the risk of disease is raised, but rather that there is a higher proportion of cases for the specific anatomic site than in the reference group. D. RESULTS D.. Florida Population and Cancer Cases D... Florida s by Percentage of County Population The distribution of Florida s as a percentage of the total population in each county in is represented as a map in Figure. Miami-Dade, Hendry, and Hardee Counties had the highest proportion of s with %, %, and % of the total population being, respectively. There was a clear south to north gradient in the proportion of s, with a greater percentage of s living in South Florida than in Central and North Florida. D.. Florida s by Percentage of Total State Population Figure is a map showing the percent distribution of total Florida s in the State by county in. A majority of s in Florida lived in South Florida, with about 7% residing in Miami-Dade, Broward, and Palm Beach counties alone. Another % resided in the counties of Orange and Hillsborough, which are located in central Florida. A smaller percentage of all Florida s lived in North Florida. D... Florida Distribution of Cancer by County - The map of Florida age-adjusted cancer rates are displayed by county and are divided into quartiles (Figure ). The age-adjusted rates are per, people and represent newly diagnosed cases between the years to, pooled to provide more stable estimates than for a single year. Data for counties with

10 less than cases are suppressed. As a result, rates for many counties in the Florida Panhandle area, which has a relatively small population, are suppressed. Counties within the highest quartile (with rates between and ) were dispersed throughout the state, although not in the South Florida counties where the majority of Florida s s lived. Counties in the third highest quartile with rates between and 9 were also distributed throughout the state with the highest number of contiguous counties located in the Central West part of Florida. Counties in the lower two quartiles were predominantly located in the Central and Southern parts of Florida. D... Percentage of New Cancer Cases by Ethnicity and Site, Figures and compare the ranking of primary sites for the year with the highest proportion of cancer diagnoses in males as compared to non- White males, and females as compared to non- White females. Rankings are in order of the top non- White cancer diagnoses. Among males, ethnic-specific cancer site rankings were similar for the top four cancers (prostate, lung and bronchus, colorectal, and head and neck). Proportionally, male s had lower cases of melanoma and bladder cancer, and slightly higher proportions of Non Hodgkin s lymphoma, liver and stomach cancer. Among non- White females, the top five cancers consisted of cancer of the breast, lung and bronchus, colorectal, and uterus, while for females the top five were cancers of the breast, colorectal, lung and bronchus, uterus, and thyroid. females also had a higher proportion of cervical cancer, but a lower proportion of melanoma than did non- White females.

11 Figure.

12 Figure. 7

13 Figure.

14 Figure. Percentage of New Male Cancer Cases by Ethnicity and Site Florida, Prostate Lung & Bronchus. 9. Colorectal Head and Neck Melanoma Bladder Non Hodgkin Kidney..7 Pancreas..7 Leukemia.. Esophogus.. Liver.. Stomach.9. Brain & Nervous.. Multiple Myeloma.. Thyroid.. Gall Bladder.7. All Others Figure. Percentage of New Female Cancer Cases by Ethnicity and Site Florida, Breast Lung & Bronchus Colorectal Uterus Non Hodkin Melanoma Ovary Pancreas Kidney Thyroid Head & Neck Leukemia Cervix Bladder Brain & Nervous Stomach Multiple Myeloma Esophogus Liver Gall Bladder All Others 9

15 D.. Trends Top Cancers Joinpoint Results The following are the results of the Joinpoint analyses of the rates of the Top Cancers over the 99- time period comparing males and females to non- White males and females by each cancer and by stage of that cancer. The overall statistical significance was set to ρ <., and a maximum of three joinpoints and four line segments was allowed for each model. While all joinpoints presented in the following graphs are statistically significant, meaning there is a significant shift in the rate of cancer at a given point in time, the slope between joinpoints or the Annual Percent Change (APC), may not be significantly different from zero, as indicated in the graphs. For each cancer, there are figures illustrating the Joinpoint analysis described in the text. These descriptions are in reference to observed changes in slopes, but not to the statistical significance of APC figures. Statistical significance of APC figures are documented in tables located on the FCDS Website and indicated with a symbol on the graph next to corresponding trend lines. Differences in APCs between groups were not evaluated. Tables of both the observed and modeled joinpoint rates are also available on the FCDS Website ( When evaluating Joinpoint analyses, decreasing trends of cancer rates are represented by downward sloping trend lines between the start of the analysis period in 99 and the end in, while upward sloping trend lines represent increasing trends in cancer rates. In some cases, there will be variable trends (e.g. initially increasing and then decreasing, or vice versa) over the entire time interval. Year specific rates reported in the text represent the observed rate, whereas Joinpoint graphs depict the predicted rate. While summaries of each cancer are described by Joinpoint trends over the entire study period, the observed rates are also included in the text to highlight the most current figures. When comparing cancer rates between s and non- Whites, health disparities can be demonstrated when either the rate of one group is decreasing while the other is increasing or remaining stable, or when the rate of decrease is different between the two groups. With regards to stage at diagnosis comparisons, health disparities would be demonstrated if there are increasing rates of persons diagnosed at distant and regional stage and/or decreasing rates of persons diagnosed at local stage since, in general, early diagnosis of cancer is associated with decreased morbidity and mortality. In particular, early diagnosis at a local stage is the anticipated goal for all screenable cancers (such as breast, colorectal, prostate, and cervical cancers). A particular issue for s is the unusually high levels of Not Otherwise Specified (NOS) cancers with regards to stage. Studies have shown that unstaged NOS cancer cases are more likely to occur among the elderly, minority populations, patients with lower levels of education and income, and those with a history of Medicaid or Medicare enrollment. 7-9 Lack of staging information suggests that patients did not receive full diagnostic evaluations, which inhibit disease management and outcome of care assessments. Therefore, constant or increasing rates of NOS cancers can be indications of health disparities. D... All Cancers a) All cancers: vs. non- White by gender The highest age-adjusted overall cancer rates in were found in non- White males ( cases/, persons), followed by males ( cases/,), then non- White females ( cases/,), and females ( cases/,). For all cancers, males had higher rates of cancer than females, regardless of ethnicity throughout the 99- time period. In general, for both s and non- Whites, after an initial increase in the first few years, there were decreasing trends in overall cancer rates over time. There was some indication of health disparities in that the decrease in cancer rates was greater for both non- White males and females compared to males and females for all cancers. b) Stage: local, regional, distant: vs. non- White by gender Again after an initial increase, there was a decreasing trend in the rates of persons diagnosed at local and regional stage which was greater for non- Whites compared to s; distant stage diagnosis rates increased for non- Whites and females over the time period. The trends of persons diagnosed as NOS decreased for all cancers for both s and non- Whites, but with a greater decrease for non- Whites.

16 Figure. and non- Joinpoint Incidence Trends, All Cancers by Sex and Stage, 9- All Sites Male 9 7 All Sites Female 9 7 All Sites Male Local Stage All Sites Female Local Stage All Sites Male Regional Stage All Sites Male Distant Stage Statistically significant Annual Percent Change at P= All Sites Female Regional Stage All Sites Female Distant Stage

17 D... Lung and Bronchus Cancer a) All lung and bronchus cancers: vs. non- White by gender Overall throughout the time period, the age-adjusted rates of lung and bronchus cancer were substantially greater for non- White males (: 9 cases/,) and females (: cases/,) compared to males (: 9 cases/,) and females (: cases/,). Initially during this time period, and non- White males and females experienced increasing trends in the rates of lung and bronchus cancer. However, by the second half of the time period, non- White and males experienced decreasing rates, while non- White and females had a substantially slower or no decrease in rates. Figure 7. and non- Joinpoint Incidence Trends, Lung & Bronchus Cancer by Sex, Lung and Bronchus Male 9 7 Lung and Bronchus Female Statistically significant Annual Percent Change at P=. b) Lung and bronchus cancer stage: local, regional, distant: vs. non- White by gender Irrespective of gender and stage, s had lower rates of lung cancer relative to non- Whites. There was a marked increase early in the time period for males and females and non- White females in diagnosis at the local stage, which had slowed for both female subpopulations and become a decreasing trend for males by the end of the time period. There were slight increases for both and non- White females at the regional stage. Non- White males had a gradual decrease in local and regional diagnosis over the time period as did males at regional stage. Diagnosis of distant stage increased slightly among all subpopulations following a brief decline among non- White males. The rates of persons diagnosed as NOS decreased for both s and non- Whites, but with a greater decrease for non- White and females.

18 Figure. vs non- Joinpoint Incidence Trends, Lung & Bronchus Cancer by Sex and Stage, 9- Lung and Bronchus Male Local Stage Lung and Bronchus Female Local Stage Lung and Bronchus Male Regional Stage Lung and Bronchus Female Regional Stage Lung and Bronchus Male Distant Stage Lung and Bronchus Female Distant Stage Statistically significant Annual Percent Change at P=.

19 D... Prostate Cancer (males only) a) All prostate cancers: vs. non- White males Overall throughout the time period, the age-adjusted rates of prostate cancer were similar for non- White males (: cases/,) compared to males (: cases/,). Both non- White and males had substantial increases in prostate cancer at the beginning of the time period, possibly associated with the increased use of PSA as a prostate cancer screening tool. However by the early 99s, substantial decreases in the prostate cancer trends occurred for both non- White and males, starting somewhat earlier for non- White males. b) Prostate cancer stage: local, regional, distant: vs. non- White males Over the time period, the rates of local prostate cancer variably increased and decreased, ultimately decreasing for both non- White and males with marked decreases in the regional and distant stage rates, possibly indicating increased early screening for this cancer. Although the rates of persons diagnosed as NOS initially increased for both s and non- Whites, by the end of the time period there were substantial decreases in these rates for both subpopulations. Figure 9. vs non- Joinpoint Incidence Trends, Prostate Cancer by Stage, Prostate 9 7 Prostate Local Stage Prostate Regional Stage Prostate Distant Stage Statistically significant Annual Percent Change at P=.

20 D... Breast Cancer (females only) a) All breast cancers: vs. non- White females Throughout the time period, the age-adjusted rates of breast cancer were substantially greater for non- White females (: cases/,) compared to females (: 97 cases/,). Over the time period, the breast cancer rate decreased for both and non- White females, with the largest decrease among non- White females. b) Breast cancer stage: local, regional, distant: vs. non- White females Over the time period, both non- White and females were decreasingly diagnosed at distant and regional stage, and increasingly diagnosed at a local stage particularly for non- White females, possibly indicating increased early screening for this cancer. The rates of persons diagnosed as NOS decreased substantially for both s and non- White females throughout the time period. Figure. vs non- Joinpoint Incidence Trends, Breast Cancer by Stage, Female Breast Female Breast Local Stage 7 9 Female Breast Regional Stage Statistically significant Annual Percent Change at P= Female Breast Distant Stage D... Colorectal Cancer a) All colorectal cancers: vs. non- White by gender The age-adjusted rates of colorectal cancer changed substantially over the time period, ultimately males had the highest rates (: cases/,) compared to non- White males (: cases/,), and females had the highest rates (: cases/,) compared to non- White females (: 7 cases/,). Over the time period, although initially both male and female s had an increase in cancer trends, there were overall decreases in colorectal cancer trends for both and non- Whites. The increase in rates for males was particularly large from eliminating the lower relative difference in rates with non- males.

21 Figure. vs non- Joinpoint Incidence Trends, Colorectal Cancer by Sex, Colorectal Male 7 Colorectal Female Statistically significant Annual Percent Change at P=. b) Colorectal stage: local, regional, distant: vs. non- White by gender Over the time period, both non- White and males and females were decreasingly diagnosed at regional stage, while the rates of local stage diagnosis mildly increased in all populations towards the end of the time period, possibly indicating a positive screening effect. females and males appeared to have increasing rates of diagnosis at distant stage, while non-s experienced decreasing distant stage diagnoses indicating a possible growing health disparity. Ultimately, the rates of persons diagnosed as NOS decreased substantially for both s and non- White males and females throughout the time period. Figure. vs non- Joinpoint Incidence Trends, Colorectal Cancer by Sex and Stage, Colorectal Male Local Stage Colorectal Male Regional Stage Statistically significant Annual Percent Change at P=. 7 Colorectal Female Local Stage Colorectal Female Regional Stage

22 9 7 Colorectal Male Distant Stage 9 7 Colorectal Female Distant Stage Statistically significant Annual Percent Change at P=. D... Bladder Cancer (including in situ) a) All bladder cancers: vs. non- White by gender Overall throughout the time period, the age adjusted rates of bladder cancer were substantially greater for non- White males (: cases/,) and females (: cases/,) compared to males (: cases/,) and females (: cases/,). Bladder cancer trends decreased among both male and female s throughout the time period. Despite some variability in trend, ultimately bladder cancer rates decreased among both male and female non- Whites. Figure. vs non- Joinpoint Incidence Trends, Bladder Cancer by Sex, 9- Bladder Male 9 7 Bladder Female Statistically significant Annual Percent Change at P=. b) Bladder cancer stage: in situ local, regional, distant: vs. non- White by gender Over the time period, both males and females were increasingly diagnosed at in situ stage with decreased diagnosis at all other stages; non- White males and females experienced a similar increasing trend at the beginning of the time period with a decrease in early diagnosis towards the end of the time period particularly for non- White females. Regional diagnosis rates decreased for all subgroups except for females. There were greater relative reductions in distant diagnoses seen among males relative to non- White males who saw no change in rates over the time period. females also experienced greater reductions in the rate of distant bladder cancer relative to non- White females. Ultimately, the rates of persons diagnosed as NOS decreased substantially for both s and non- White males and females throughout the time period. 7

23 Figure. vs non- Joinpoint Incidence Trends, Bladder Cancer by Sex and Stage, 9- Bladder Male In Situ 9 7 Bladder Female In Situ Bladder Male Local Stage 9 7 Bladder Female Local Stage Bladder Male Regional Stage Bladder Male Distant Stage Bladder Female Regional Stage Bladder Female Distant Stage Statistically significant Annual Percent Change at P=.

24 D..7. Head and Neck Cancer a) All head and neck cancers: vs. non- White by gender Overall throughout the time period, the age-adjusted rates of head and neck cancer were greater for non- White males (: 9 cases/,) and females (: cases/,) compared to males (: cases/,) and females (: 7 cases/,). Over the time period, there was an overall decrease in head and neck cancer trends for both and non- Whites, with a dramatic reduction in rates noted for males which started around 99. Figure. vs non- Joinpoint Incidence Trends, Head & Neck Cancer by Sex, 9- Head and Neck Male Head and Neck Female Statistically significant Annual Percent Change at P=. b) Head and neck cancer stage: local, regional, distant: vs. non- White by gender Over the time period, the overall trend was decreasing diagnosis at local stage; additionally non- White females and males and females were decreasingly diagnosed at regional stage. However, non- White and males had increasing rates of distant stage diagnosis over the time period. Ultimately, the rates of persons diagnosed as NOS decreased for both s and non- White males and females throughout the time period, particularly for males. Figure. vs non- Joinpoint Incidence Trends, Head & Neck Cancer by Sex and Stage, Head and Neck Male Local Stage 9 7 Head and Neck Female Local Stage Statistically significant Annual Percent Change at P=. 9

25 Head and Neck Male Regional Stage Head and Neck Male Distant Stage Head and Neck Female Regional Stage Head and Neck Female Distant Stage Statistically significant Annual Percent Change at P=. D... Non Hodgkin s Lymphoma (NHL) a) All non Hodgkin s lymphoma cancers: vs. non- White by gender Overall, the age-adjusted rates of non Hodgkin s lymphoma cancer were variable throughout the time period. Ultimately, however the rates were similar for Males (: cases/,) and non- White males (: cases/,), as well as for non- White females (: cases/,) and females (: cases/,). Over the time period, males had decreasing trends in Non Hodgkin s lymphoma, while after initial increasing trends, non- White males and females also ultimately had decreasing rates of non Hodgkin s lymphoma. However, females experienced increasing trends throughout the time period. Figure 7. vs non- Joinpoint Incidence Trends, NHL Cancer by Sex, Non Hodgkin's Lymphoma Male Statisticallysignificant Annual PercentChangeatP=. 9 7 Non Hodgkin's Lymphoma Female b) Non

26 Hodgkin s lymphoma stage: local, regional, distant: vs. non- White by gender Over the time period, both non- White and males and females were increasingly diagnosed at local and regional stages. While initially there was an increase in distant stage Non Hodgkin s lymphoma for non- White males, there is a general decrease for both non- White and males throughout the time period. and non- White females were increasingly diagnosed at distant stage during this time period. The trends of persons diagnosed as NOS decreased for both s and non- White males and females throughout the time period, particularly for non- White males and females. Figure. vs non- Joinpoint Incidence Trends, NHL Cancer by Sex and Stage, Non Hodgkin's Lymphoma Male Local Stage.... Non Hodgkin's Lymphoma Female Local Stage Non Hodgkin's Lymphoma Male Regional Stage Non Hodgkin's Lymphoma Female Regional Stage Non Hodgkin's Lymphoma Male Distant Stage Non Hodgkin's Lymphoma Female Distant Stage Statistically significant Annual Percent Change at P=.

27 D..9. Melanoma a) All melanoma cancers: vs. non- White by gender The age-adjusted rates of melanoma cancer were substantially greater for non- White males (: 7 cases/,) and females (: 7 cases/,) compared to males (: cases/,) and females (: cases/,). Over the time period, trends in melanoma increased substantially for male and female non- Whites, while rates for males and females were generally stable. Figure 9. vs non- Joinpoint Incidence Trends, Melanoma by Sex, 9- Melanoma Male Melanoma Female Statistically significant Annual Percent Change at P=. b) Melanoma stage: local, regional, distant: vs. non- White by gender Over the time period, both non- White females and males were increasingly diagnosed at a local and regional stage, possibly indicating increased early screening for this cancer. females had too small numbers to establish trends for distant stage, however, trends for all other groups were generally stable. The trends of persons diagnosed as NOS decreased for males and for non- White males and females throughout the time period, while females experienced a slight increasing trend. Figure. vs non- Joinpoint Incidence Trends, Melanoma by Sex and Stage, 9- Melanoma Male Local Stage Melanoma Female Local Stage Statistically significant Annual Percent Change at P=.

28 Melanoma Male Regional Stage Melanoma Male Distant Stage Melanoma Female Regional Stage * Melanoma Female Distant Stage Statistically significant Annual Percent Change at P=. *Linear model with missing values * D... Ovarian Cancer (females only) a) All ovarian cancers: vs. non- White females The age-adjusted rates of ovarian cancer were similar between non- White females (: cases/,) and females (: cases/,). Both and non- White females had decreasing trends in ovarian cancer. b) Ovarian cancer stage: local, regional, distant: vs. non- White females Over the time period, both non- White and females were decreasingly diagnosed at local stage and increasingly diagnosed at regional stage with a decrease in diagnosis at distant stage for both subpopulations. The trends of persons diagnosed as NOS decreased for both and non- White females throughout the time period, with a greater decrease among non- White females. Figure. vs non- Joinpoint Incidence Trends, Ovarian Cancer by Stage, Ovary Statistically significant Annual Percent Change at P=..... Ovary Local Stage

29 .... Ovary Regional Stage Statistically significant Annual Percent Change at P= Ovary Distant Stage D... Cervical Cancer (females only) a) All cervical cancers: vs. non- White females The age-adjusted rates of cervical cancer were slightly higher for females (: cases/,) compared to non- White females (: 9 cases/,). Over the time period, both and non- White females experienced decreasing trends of cervical cancer. b) Cervical cancer stage: local, regional, distant: vs. non- White females Over the time period, both non- White and females were decreasingly diagnosed at local, regional and distant stages with the greatest decreases seen among females. The trends of persons diagnosed as NOS decreased for both and non- White females throughout the time period, with a greater decrease among non- White females. Figure. vs non- Joinpoint Incidence Trends, Cervical Cancer by Stage, Cervix Cervix Local..... Cervix Regional Statistically significant Annual Percent Change at P= Cervix Distant Stage

30 D... Stomach Cancer a) All stomach cancers: vs. non- White by gender Overall throughout the time period, the age-adjusted rates of stomach cancer were higher among males (: cases/,) compared to non- White males (: 7 cases/,) and females (: cases/,) compared to non- White females (: cases/,). Over the time period, both male and female s and non- White females had a decreasing trend in stomach cancer rates. Despite an initial increasing trend in non- White males, ultimately there was a decreasing trend by the end of the time period. For both males and females, the rate of decline in stomach cancer incidence was similar in s versus non- Whites over the 99- time period. Figure. vs non- Joinpoint Incidence Trends, Stomach Cancer by Sex, Stomach Male 9 7 Stomach Female Statistically significant Annual Percent Change at P=. b) Stomach cancer stage: local, regional, distant: vs. non- White by gender Over the time period, there was a slight increasing trend in local stage rates among s and a generally stable trend among non- Whites. Rates for regional stage decreased for all subpopulations and decreases in distant stage rates are observed among males and non- White females. The trends of persons diagnosed as NOS decreased for both and non- White males and females throughout the time period. Figure. vs non- Joinpoint Incidence Trends, Stomach Cancer by Sex and Stage, Stomach Male Local Stage Stomach Female Local Stage Statistically significant Annual Percent Change at P=.

31 .... Stomach Male Regional Stage.... Stomach Female Regional Stage.... Stomach Male Distant Stage.... Stomach Female Distant Stage Statistically significant Annual Percent Change at P=. D... Liver Cancer a) All liver cancers: vs. non- White by gender Overall throughout the time period, the age-adjusted rates of liver cancer were higher among males (: cases/,) compared to non- White males (: cases/,) and females (: cases/,) compared to non- White females (: cases/,). Over the time period, there were increasing trends in the liver cancer rates for both and non- Whites. Figure. vs non- Joinpoint Incidence Trends, Liver Cancer by Sex, Liver Male..... Liver Female Statistically significant Annual Percent Change at P=.

32 b) Liver cancer stage: local, regional, distant: vs. non- White by gender Over the time period, both non- White and males and females were increasingly diagnosed at a local and regional stage. Trends in rates for distant stage diagnoses were generally stable among females and were slightly decreasing among males. The trends of persons diagnosed as NOS decreased for both and non- White males and females throughout the time period. Figure. vs non- Joinpoint Incidence Trends, Liver Cancer by Sex and Stage, Liver Male Local Stage..... Liver Female Local Stage.... Liver Male Regional Stage.... Liver Female Regional Stage Liver Male Distant Stage Liver Female Distant Stage Statistically significant Annual Percent Change at P=. 7

33 D... Gall Bladder Cancer a) All gall bladder cancers: vs. non- White by gender Ethnic differences in the rates of gall bladder cancer diminished over time. Overall, the age-adjusted rates of gall bladder cancer were similar between males (:./,) and non- White males (:.7 cases/,), as well as females (: cases/,) and non- White females (: cases/,). Over the time period, there was a decreasing trend in gall bladder cancer for both males and females in contrast to the rates for non- Whites which remained stable through the study period. Figure 7. vs non- Joinpoint Incidence Trends, Gall Bladder Cancer by Sex, Gall Bladder Male.... Gall Bladder Female Statistically significant Annual Percent Change at P=. b) Gall bladder cancer stage: local, regional, distant: vs. non- White by gender Non- White males and females had relatively stable trends of diagnosis at local, regional, and distant stages. females experienced reductions in the incidence of gall bladder cancer at all stages, while males experienced incidence reductions for local and regional stage only. The numbers of cancers were relatively small, leading to unstable trends for s by stage. Figure. vs non- Joinpoint Incidence Trends, Gall Bladder Cancer by Sex and Stage, Gall Bladder Male Local Stage Statistically significant Annual Percent Change at P= *Linear model with some missing values Gall Bladder Female Local Stage *

34 Gall Bladder Male Regional Stage * Gall Bladder Male Distant Stage Gall Bladder Female Regional Stage Gall Bladder Female Distant Stage Statistically significant Annual Percent Change at P=. *Linear model with some missing values D.. Percentage Distributions for All Cancer D... s Versus non- Whites by Gender The total number of cancers (with females having the fewest and non- White males have the largest number of cancers) was substantially greater for the non- Whites compared to the s, as would be expected based on the number of non- Whites and s residing the Florida population as well as their individual cancer rates: female cases (7,), male cases (,), non- White female cases (9,7), and non- White male cases (79,). The percentage distribution of the major cancers by gender and by ethnicity is displayed in both tabular and graphic form (Table and Figure 9). males and females had fewer of the tobacco-related cancers (i.e. lung, head and neck, and bladder) than non- White males and females, respectively (except for head and neck in males). females had a higher percentage of cervical, colorectal, gall bladder, liver, non Hodgkin s lymphoma, and stomach cancers, while non- White females had a higher percentage of melanoma; there were basically equal percentages of breast and ovarian cancers. males had a higher percentage of colorectal, gall bladder, liver, non Hodgkin s lymphoma, and prostate and stomach cancers, while non- White males had a higher percentage of melanoma. females had the greatest percentage of cancers from other sites, while non- White males had the fewest. 9

35 Figure 9. vs. non- Percent Cancer Distribution by Site and Sex, 99- Lung

36 Table. Percent Distribution of Major Cancers by Ethnicity and Sex Cancer type (%) Female Male non White non White Lung Prostate 9.9. Breast (female).. Colorectal...9. Bladder Head & Neck...7. Non Hodgkin s Lymphoma...7. Melanoma.... Ovarian..7 Cervical.. Stomach.... Liver.... Gall bladder.... All other sites All Cancers (N) 7, 9,7, 79, D... Proportional Distribution of Cancers by Subpopulation and Gender The distribution of the selected cancers for each subpopulation was compared to Non- Whites by gender using the proportional incidence ratio methodology. The proportions that make up these ratios are presented by subgroups shown in figure. Among both females and males, in general the proportions of the relatively rare cancers (i.e. stomach, liver, and gall bladder) were substantially elevated compared to Non- Whites. In addition, the proportion of cervical cancer was elevated among all the females subpopulations, while there was an increased proportion of non Hodgkin s lymphoma and of all other cancers for many of the subpopulations for both males and females. Among the subpopulations by gender, South and Central American females had the highest proportions of stomach, gall bladder, cervix, and all other cancers, while Mexican females had the highest proportion of liver cancer; Puerto Rican females the highest non Hodgkin s lymphoma; and Cuban and not otherwise specified females had the highest proportion of colorectal cancer. Somewhat similar patterns were seen for the subpopulation males: South and Central American males had the highest proportions of stomach, gallbladder, non Hodgkin s lymphoma, and all other cancers; while Puerto Rican males had the highest proportion of liver cancer, and not otherwise specified males had the highest colorectal and prostate cancer proportions. Figure. Percent Cancer Site Distribution for Sub-Groups, by Site, and Sex, 99-